Population dynamics and epibiont associations of hermit crabs (Crustacea: Decapoda: Paguroidea on Dog Island, Florida

8 Pages·2003·5.9 MB·English

by  F Sandford

MemoirsofMuseumVictoria60(1):45-52(2003)
ISSN 1447-2546(Print) 1447-2554(On-line)
http://www.museum.vic.gov.au/memoirs
Population dynamics and epibiont associations ofhermit crabs (Crustacea:
Decapoda: Paguroidea) on Dog Island, Florida
Floyd Sandford
DepartmentofBiology,CoeCollege,CedarRapids,Iowa52402,USA([email protected])
Abstract Sandford,F.2003.Populationdynamicsandepibiontassociationsofhermitcrabs(Cmstacea:Decapoda;Paguroidea)on
DogIsland,Florida.In:Lemaitre,R.,andFudge,C.C.(eds).BiologyoftheAnomura.Proceedingsofasymposiumat
theFifthInternationalCrustaceanCongress,Melbourne,Australia,9-13July2001.MemoirsofMuseumVictoria60(1):
45-52.
Periodicbelttransects anddaily shoreline surveysinJanuary andinJune-Julyovera 10-yearperiod(1992-2001)
wereusedtostudythesevenspeciesofhermitcrabsmostcommonintheupperintertidalzoneofalowenergybaywith
a960mshorelineattheeastendofDogIsland,Florida,andtheirassociationwiththreecommonshellepibiontsinthe
area:theFloridahermit-crabsponge,Pseudospongosoritessuberitoides,thecloakanemone,Calliactistricolor,andthe
hydrozoan Hydractinia echinata. Of 15,052 hermit crabs sampled, Pagurus longicarpus, Pagurus pollicaris, and
PagurusimpressuswereprevalentinJanuary(88%ofallanimals)andClibanariusvittatusdominatedinthesummer
(86%ofallanimals).Thefollowingassociationswerehighlysignificant:P impressuswithP suberitoides,Ppollicaris
withH. echinata, andPetrochirusdiogenes with C. tricolor. C. vittatusrarelyhad anemones, andP impressus were
neverfoundinshellswithH.echinataandshowedasignificanttendency,whetherinspongeorshellshelters,tobecome
strandedcomparedtootherpaguridspecies. Hermitcrabspongeswerecommonlyusedas sheltersbyonlytwoofthe
hermitcrab species, P. impressus andPaguristeshummi. Togetherthese two species accountedfor99% ofthe 1077
animalsfoundinsponges.Hermitcrabspongesvariedyearlyinabundancefromplentifultouncommon,andnearlyhalf
(1030of2107,or49%)wereempty.
Keywords Cmstacea,Anomura,hermitcrab,populationdynamics,epibiont
Introduction and Kelly-Borges, 1997). The sponge, Pseudospongosorites
StudiesofhermitcrabsonDogIsland,Florida,abarrierisland suberitoidesDiazetal., 1993 (reclassifiedbyMcCormackand
in the north-eastern comer ofthe GulfofMexico, over a 10 Kelly, 2002) is one ofthe compact and colourful hermit-crab
year period have shown the seven most common species fsapmoinlgyes,Subaeruitniidqauee) grrepoourpteodffsrpoomngmeasny(orddiefrferHeandtrolmoceartiidoan,s
encounteredinintertidalwaterssurroundingtheislandtobeas
worldwide (Sandford and Kelly-Borges, 1997). Hermit-crab
follows: Pagurus longicarpus Say, 1817, Pagurus pollicaris
sponges typically colonise a living or dead gastropod shell,
Say, 1817, Pagurus impressus (Benedict, 1892), Pagurus
althoughother substrates (e.g. othermollusc shells, inanimate
stimpsoni (A. Milne Edwards and Bouvier, 1893), Paguristes
objects such as floating docks or wharf pilings) are used
hummi Wass, 1955, Petrochirus diogenes (Linnaeus, 1758), (SandfordandBrown,1997).Thespongeeventuallyovergrows
and Clibanarius vittatus (Bose, 1802). Pagurus longicarpus the shell which becomes increasingly more deeply enclosed
and C. vittatusareintertidal species(StrasserandPrice, 1999;
within the sponge mass. Ifthe shell is occupied by a hermit
W. Herrnkind, pers. comm.). Clibanarius vittatus is a hardy crab,thecrabeventuallyvacatestheshellandoccupiesacham-
speciesthatoftenenterstheupperintertidalandcanspenddays ber withinthe spongebody, moving aboutinthe sponge with
outofwater(Rudloe, 1984). Paguruspollicarisis primarily a onlyitsanteriorendvisiblethroughanopeningmaintainedby
shallow subtidaltolowerintertidal species, andthe otherfour
the crab. Nearly allhermit-crab sponges worldwide are found
aretypicallyshallowtodeepsubtidal(StrasserandPrice, 1999; in deep water, typically recovered by dredging, but on Dog
Herrnkind,pers.comm.),althoughP. impressusmovesintothe Island such sponges are abundant in the intertidal zone inthe
pi1ny9t9e7gr)at.isdtAarlsozpioosdntesmhdeeulrfliosrnbgmuotthsettwwhoie,nrtPme.irtimmcoprnratbessh,ssuas(llSaasnneddvfePonrdshpuaemncmidieBs,rococowcmnu-,- fwsirpnootnmegresspmononengaterhssshionwrteho.eInanvatPi.hleaibllmaepborresahstesloulrssy,,iPanbaisnmppdoronengsiesnugsshtoelhfteteernsspsowbnirgtiecnsgh
monlyusespongeshelters(Wass, 1955;Wells, 1969; Sandford (Sandford, 1995).Becausespongesaremoreeasilyaffectedby
46 F. Sandford
waveactionthanareshells,hermitcrabsinspongesnearshore- the first year, all analyses of hermit crab populations were
line are more likely to be affected by waves or beached basedondatafromthelastnineyearsofthestudy.
by receding tides. During the month of January many
TestingforshelterpreferenceandfidelityinPagurusimpressus.
sponges, bothempty and occupiedbyhermitcrabs, are found
Previousstudiesinthelaboratory(Sandford, 1995)showedthat
nearthe shoreline,orstranded anddryingonshore (Sandford,
individuals ofP. impressus prefer shell shelters. To determine
1995).
Twoothercommonlyencounteredshellepibiontsinthearea shelterpreference and shelterfidelity forrecentlyfieldcaught
animals, shell/sponge switching experiments were conducted
are the cloak anemone, Calliactis tricolor (Phylum Cnidaria, on Dog Island for three years, January 1994-1996. Shelter
Class Anthozoa) and the “snail fur” hydrozoan Hydractinia
fidelity wasmeasuredbythetendencyofanimalstoremainin
echinata (PhylumCnidaria, ClassHydrozoa).
theiroriginal shelters overthe course ofup to a 3 daytesting
To determine annual and seasonal population changes for
andobservationperiod.
thehermitcrabspeciesandtocheckforspecificityinthethree
Pagurus impressus in shell or sponge shelters collected
mentioned epibiont associations, a study was conducted on
DogIslandfor 10consecutiveyears(1992-2001). during shoreline and belt transect surveys were isolated, then
testedwithin48-72hofcapture.Of209animalscollectedand
Methods used for testing, 126 were in sponges and 83 in shells ofthe
gastropod Strombus alatus. All werejuveniles or small adults
Study site. Studies were conducted on Dog Island, St George withchela4.5-10.0mmlong.Tocontrolforshelltypeprefer-
Sound,Florida(29°49.30"N,84°34.30'W)annuallysince1992, ence, only animals found in S. alatus shells, a preferred shell
duringthemonthofJanuary(1992-2001)andduringJune-July forP. impressus(Table4),wereusedintesting.Emptysponges
for 4 years (1993, 1996, 1998, and 1999). Dog Island is the andemptyStrombusalatusshellswerealsocollected.Allshells
easternmostofachainofbarrierislandsboundingthesouthern andspongesusedas shelterchoicesinthetestswerefieldcol-
perimeterofApalachicolaBayandStGeorgeSoundintheNE lected, either empty orpreviously occupied by individuals of
Gulf of Mexico. It is a true barrier island, made of uncon- P. impressus.Allemptyspongeswerehealthy(i.e.goodcolour
solidatedsandoverlyingtheSW-dippinglimestonebedrockof withno bleaching and compact with adense, non-flabby tex-
theFloridaplatform.Theislandliesabout6kmoffshorefrom ture) and all empty S. alatus shells were undamaged and
theN. Floridapanhandle, andis 11 kmlong and 1.2kmatits epibiont-free.
widest. All transect work was done in the intertidal zone ofa
north-facingbayattheeastendoftheisland.Thebayhasa960 Initial Test. Each animal was placed between an empty
m shoreline and a sandy and sand/mud bottom with scattered S. alatus shell and an empty sponge at the center ofa round
seagrassbedsofmanateegrass,Syringodiumfdiformeandtur- (diameter 11.2cm)plasticcontainer,witha1cmdepthofsand
tlegrass Thallasiatestudinum.Waterdepthrangedfrom0.3to on the bottomand a 5 cmdeep watercolumn above the sub-
1.3 m atmeanlowtide.At 1500 hinJanuary watertempera- stratesurface.Watertemperaturerangedwithoutsidetempera-
tures ranged from 12.5-18.5°C (normally 14-17°). January ture,from 12°to 19°C.Shelterchoiceswereinthecenterofthe
2001 was atypical with temperatures as low as U°C. Work chamber 1 cmapartwiththeiroppositesidesanequaldistance
consisted of periodic belt transect surveys, daily shoreline from the walls ofthe chamber. Shelters were positioned with
surveys,andshell/spongeswitchingexperiments. apertures or openings up and new sea water added after each
trial.Thechoiceshelterwasofequalorslightlylargersizethan
Shoreline surveying. Shoreline surveys were conducted once thatoccupiedbythecrab.Thetestlasted30minutes.Animals
(0700-1000 h), and often twice (1400-1700 h), daily. All that switchedinto anew shelterwerenoted, thenreleasedand
sponges andmhermit crabs stranded on shore or in the water not tested again. It was noted whether the animal switched
within 0.5 from shoreline were collemcted and identified. short-term (remained in the new shelter for <3 min) or long-
Surveysincludedtheshorelineofthe960 lowenergybe—ach term (remained in the new shelter for >3 min; crabs that
ofthebay studymsite, inadditiontotwo adjoining shores a remainedinthenew shelterfor>3 minrarelyreturnedtotheir
cSoonutnidguuopusto70th0einsltetreotfchaotfildaolwsealntermgayrsbhayabneda2c,h3o00nmStGofeohriggeh originalshelter).
energyGulfbeachattheextremeeastendoftheisland-atotal Follow-up test. Any animal that did not switch into a new
of3,960mofshoreline.ForthemonthsofJanuary 1992-2001 shelter during the initial test was immediately returned to its
the surveys totalled 140 days and 338 h survey time; the originalcontaineralongwiththetwoemptychoicesheltersand
summersurveystotalled37daysand59hofsurveytime. observed periodically for72 h.All animals that switched into
new shelterswerenoted; forsuchanimalsthetestwasconsid-
Belttransectsurveying.Inthebelttransectsurveysfromoneto eredendediftheanimalremainedinthenewshelterfor>24h.
fourpersonsmovedfromoneendofthebaytotheother, col- DatawasanalysedforsignificanceusingMinitabstatistical
lecting all hermit crabs stranded omn shore and from shoreline software.
intotheintertidalzonetowithin8 fromshoreline.Allhermit
crabswereidentifiedtospecies,shelltypeandpresenceofepi-
m Results
biontsnoted,thenreleased200 awayinStGeorgeSound.All
samplingwasdoneonarisingtideandusuallyfroml600-1800 Sponge abundance. Sponges varied in abundance by season
h. Sixty-sixtransectsweretaken: 46fromJanuary 1993-2001, andfromyearto year.Atotalof2107 sponges were collected
and20inthesummer.Becausenotransectsamplingwasdone intenyears. Spongeswerecommoninthewinter(99.3%ofall
Populationdynamicsandepibiontassociationsofhermitcrabs 47
specimens) and uncommoninthe summer, and approximately Table 1. Hermit crab sponge abundance on Dog Island, Florida, by
half were without a hermit crab occupant (Table 1). Sponge season,overaten-yearperiod(1992-2001).
abundanceinJanuaryvariedonanannualbasisandinarough-
ly cyclical way: sponges were abundant in 1994, common in Timeofyear Empty Withhermitcrabs Total
1992 and 1996, less commonin 1993, 1995, 1997, 1999, and
2000,anduncommonin 1998and2001.In 1992,thefirstyear WSiunmtmeerr(J(a1n99139,921-92906,01) 1017 1075 2092
ofthestudy,hermitcrabspongeswerenumerousand358were 1998-1999) 13 2 15
collected; 197 (55%) were occupied by a hermit crab, all of
which were Pagurus impressus. But hermit crabs in sponges Total 1030 1077 2107
are more easily stranded on shore than those in shells andno
transect sampling was done in 1992. To control for sampling
194 were insponges (42.7%) and 260 were in shells (57.3%).
bias, all analyses of associations between hermit crabs and
Individuals of P. impressus used 15 shell types, but three,
sponges were restricted to the data gathered by both transect
Strombus alatus, Cantharus cancellarius, and Busycon con-
andshorelinesamplingfrom 1993to2001. trarium, were used by 72% of all individuals (Table 4).
HermitcrabsofDogIsland.Atotalof15,052hermitcrabs,rep- Becausenearlyallhermitcrab-occupiedsponges,whethercol-
resenting seven species, were surveyed over the ten year lectedinwateroronshore,containedindividualsofP. impres-
period. The four most common species {P. impressus, susandbecausethefrequencywithwhichemptyspongeswere
P poUicaris, P. longicarpus, and C. vittatus) comprised over foundinJanuaryvariedannually(from87%in 1997to 19%in
98% of all hermit crabs sampled annually (Table 2). All 2000) the data for 1993-2001 were analyzed to see if the
species, withtheexceptionofC. vittatus, weremostprevalent numberofempty spongessampledcorrelatedwiththenumber
during the winter. In January individuals of three species, ofP. impressussurveyed.Poorcorrelationswerefoundforboth
Pagurus impressus, P longicarpus, and P poUicaris, consti- (i) thetotalnumberofP. impressus surveyed(inbothsponges
tuted 96% ofall animals sampled; the most prevalent species and shells) (R2=0.44) and (ii) only theP. impressusfoundin
wastheintertidalspeciesP longicarpuswhichcomprised41% shells (R2=0.27).
of all animals sampled (Table 2). The typically subtidal
species P diogenes, P hummi, and P. stimpsoni occurred AHsysdorcaicattiinoina eocfhifnoautar.hFeorrmitthecfroaubr msopsectiecsomwmitohnthheermhiytdrcoraibd
sporadically. Inthe summer, individuals ofC. vittatus domin-
species, 13,995 individuals were in shells, and ofthese, 1343
ated the intertidal zone, along with some individuals of
P longicarpus and P poUicaris, whereas the other species (9.6%) were covered with hydroids of Hydractinia echinata
(Table 5). The associationofthehydroid withthe fourhermit
wererare(Table2). crab species was dramatically non-random. No individuals of
Association ofhermit crab species with sponge shelters. Two P. impressus or of C. vittatus were associated with hydroids,
species,PagurusimpressusandPaguristeshummi, werecom- whereas66%ofP.poUicarisand8%ofP. longicarpuswerein
monly found in sponge shelters. Individuals of three other hydroid-covered shells (Table 5). The differences betweenthe
species, P. longicarpus, P. poUicaris, and C. vittatus, rarely three pagurids were all highly significant (two-sided tests for
used sponge shelters (< 0.5%) (Table 3). Over half of all P. the equality oftwo proportions: P. poUicaris vs P. impressus,
impressus and nearly a third of all P. hummi found were in Z=55.5,P<0.001;PpoUicarisvsP. longicarpus,Z=44.9,P
sponges. The differenceinuse ofsponges as shelters between <0.001;P. longicarpusvsP. impressus,Z= 17.6,P<0.001).
P. impressusandP. hummicomparedtotheother3specieswas Hydroids ofH. echinata werefoundon allgastropod shell
highlysignificant{yf=9608,df=2,P<0.001).Althoughindi- typescommonlyusedbyhermitcrabsonDogIsland.However,
viduals of both P. impressus and P. hummi were found in to checkforthe possible effects ofshell substrate onhydroid
sponges, the greaterassociation with sponges byP. impressus growth, the association ofH. echinata withonly shells ofthe
was significant(twosidedtestforequalityoftwoproportions, ribbedwhelk,Cantharuscancellarius, wasexamined(Table6).
Z = 5.14, P < 0.001). Individuals of P. impressus showed a Younger animals of all three pagurid species commonly use
highlysignificantassociationwithspongeshelterscomparedto Cantharus cancellarius shells. This was especially true for
all other species and to P. poUicaris, the third most common individuals ofP. poUicaris as over half(56%) ofall animals
species using sponges (Table 3). Of all 1494 P. impressus sampledwereinCantharusshells,and92%ofthesewerecov-
collected from 1993-2001, most (56%) were in sponges, eredbyH. echinata. Cantharus shells were more likely to be
compared to only 7 of 1,621 P. poUicaris found in sponges coveredbyhydroidsthanothershelltypesandaswasthecase
(two sided test for equality of two proportions, Z = 42.6, with all shells (Table 5), the associations ofthe three pagurid
P<0.001). species in hydroid-covered Cantharus shells are all signifi-
Mostofthe 1,077 occupiedspongesinthe study contained cantlydifferentfromone another (two-sidedtestsforequality
individuals of Pagurus impressus and most of the sponges oftwo proportions: P. poUicaris vs P. impressus, Z = 102.3,
collected in the study were found stranded on shore, but a P<0.001 ;P.poUicarisvsP. longicarpus,Z=47.1,P<0.001;
noticeable association ofP. impressus with sponge shelters is P. longicarpusvsP. impressus, Z = 15.3,P<0.001). Pagurus
also evident for all animals collected in water during the belt poUicarisis commonly associated withhydroids, P. longicar-
transectsurveys (Table4).Atotalof454P. impressusindivid- pussignificantlylessso,andP. impressusneverusesCantharus
uals were collected in the surveys from 1993-2001. Ofthese shellswithhydroids(Table6).
48 F. Sandford
Table2.Summaryofhermitcrabscollectedbybelttransectandshorelinesurveysfornineyears, 1993-2001,byseason.
Winter Summer
%of %ofall %of %ofall
Species Total Number species animals Number species animals
Pagurusimpressus 1,494 1,489 99.7 27.6 5 0.3 <0.1
Paguruspollicaris 1,621 1,471 90.7 27.3 150 9.3 <1.5
Paguruslongicarpus 3,399 2,203 64.8 40.9 1,196 35.2 12.4
Clibanariusvittatus 8,323 13 0.2 0.2 8,310 99.8 86.0
Petrochirusdiogenes 39 38 97.4 0.7 1 2.6 <0.1
Paguristeshummi 117 116 99.1 2.2 1 0.9 <0.1
Pagurusstimpsoni 54 54 100.0 1.0 0 0.0 0
unidentified 5 5 100.0 0.1 0 0.0 0
Totals 15,052 5,389 100.0 9,663 100.0
Table3.AssociationofPagurusimpressusandPaguristeshummi,two Table6.AssociationofthreepaguridswithshellsofCantharuscan-
species commonly found in sponges, with five other hermit crab cellarius,withorwithoutthehydroid,Hydractiniaechinata, onDog
species that rarely or never used sponge shelters for nine years Island,Floridafornineyears, 1993-2001.
(1993-2001).
Species Numbers %in % with
Species Numbers %insponges inShells Cantharus H. echinata
Pagurusimpressus 1,494 55.5 Pagurusimpressus 662 23.3 0.0
Paguristeshummi 117 32.5 Paguruspollicaris 1,614 56.4 92.0
Paguruspollicaris 1,621 0.4 Paguruslongicarpus 3,397 28.8 19.2
Paguruslongicarpus 3,399 0.1
Clibanariusvittatus 8,323 <0.1 Total 5,673
Petrochirusdiogenes 39 0
Pagurusstimpsoni 54 0 Table7.Associationoffivehermitcrab speciesfoundinshellswith
Unidentified 5 0 thecloakanemone,Calliactistricolor,onDogIsland,Florida,fornine
years, 1993-2001.
Total 15,052 5.85
Species Numbersinshells %withC. tricolor
Table4. Sheltersoccupiedby 454individualsofPagurus impressus
collectedin66belttransectsurveys, 1993-2001. Pagurusimpressus 662 1.3
Paguruspollicaris 1,614 5.5
Shelter Numbers %total %of shellspecies Paguruslongicarpus 3,397 0
Clibanariusvittatus 8,322 0.1
insponges 194 42.7 Petrochirusdiogenes 39 66.7
inshells; 260 57.3
Strombusalatus 76 29.2 Total 14,034 1.2
Cantharuscancellarius 62 23.9
Busyconcontrarium 50 19.2
Busyconspiratum 20 7.7
Chicoreusdilectus 12 4.6
others(10spp.) 40 15.4 Table 8. Test of shelter preference for 126 individuals ofPagurus
impressuscapturedinsponges.Initialtestlasting30minutes,check-
Total 454 100.0 100.0 ing whetherthe crab switchedinto another sponge, into aStrombus
alatusshell,intobothspongeandshell,ordidnotswitch.Two-tailed
Table5.Associationofthefourmostcommonhermitcrabspecies(for binomialprobabilitytest,Hq:Psponge=0-5,P=0.0004,significant.
allindividualsfoundinshells)withthehydroidHydractiniaechinata
onDogIsland,Florida,fornineyears, 1993-2001 Switchedfromspongeinto: Numbers %
Species Numbersinshells Percentwith Sponge 9 7.1
H. echinata Shell 32 25.4
Both 4 3.2
Pagurusimpressus 622 0 noswitch 81 64.3
Paguruspollicaris 1,614 65.6
Paguruslongicarpus 3,397 8.4 Total 126 100.0
Clibanariusvittatus 8,322 0
Total 13,995 9.6
Populationdynamicsandepibiontassociationsofhermitcrabs 49
Table 9. Test of shelter preference for 83 individuals ofPagurus were surveyed, 1,074 in the water and 54 (4.8 %) on shore.
impressusinStrombusalatusshells.Initialtestlasting30minutestest- Overthesameperiod283R impressusinshellsweresurveyed,
ingwhetherthehermitcrabswitchedfromitsshellintoasponge,into 126 in the water and 157 (55.5%) on shore. Compared to
sawniottchhe.rTSwtor-otmabiulsedalbaitnusomsihaelll,prionbtaobibloitthystpeostn,geHqa!ndPsshheeilil,=or0-d5,idPno=t fP.icpaonltlliycagrriesa,teirndliivkiedluiahlosodofofP.beicmoprmeisnsgusstirnansdheeldlsonshshoowrea(stiwgon--
0.0006,significant. sidedtestforequalityoftwoproportions,Z= 16.8,P<0.001),
Switchedfromspongeinto: Number % and areeven more likely tobe found on shore than the hardy
upperintertidalspecies C. vittatus.
sponge 1 1.2 ShelterpreferenceandfidelityforPagurusimpressus.(i).Initial
shell 15 18.1
andfollow-uptestforanimalsinsponges. Intheinitialchoice
both 1 1.2
noswitch 66 79.5 test lasting 30 minutes, of 126 P. impressus in sponges, 45
animals(36%) switchedintoadifferentshelter, showingasig-
Total 83 100.0 nificantpreferenceforshells(P=0.0004)(Table8).Inthefol-
low-uptestforthe 81 animalsthatdidnotswitchintheinitial
test, 43 (53%) switched shelters; of these, 42 of43 (97.7%)
Associationoffivehermitcrabspecieswiththecloakanemone, selected the shell (P < 0.0001). In summary, 88 of the 126
Calliactistricolor.Theindividualsofallfourspeciespresentin (70%)animalscapturedinspongesswitchedintoanewshelter,
shells that were checked for H. echinata (Table 5) were also showingahighlysignificantpreferenceforshellsoversponges.
examinedforthepresence ofthe cloakanemone, Calliactistri- (ii). Initial and follow-up test for animals in shells. In the
color. Datafor afifth species, Petrochirus diogenes, was also initialchoicetest,of83animalscapturedinS. alatusshells, 17
recorded; this species never used sponges (Table 3) or shells animals (20%) switchedinto anew shelter, showing a signifi-
coonveeorredmowriethaHn.emeocnheinsat(aT,abbluet7)w.aAs tcootamlmoofnl1y4,0f3o4unidndciavrirdyuialnsg Icnanttheprfeoflelroewn-cuepftoerstotohferthSe.6a6laatnuismsahleslltsha(tPd=id0.n0o0t0s6w)i(tTcahbilnet9h)e.
ofthe five species was collected in shells, of which 173 had initialtest,25(38%)switchedintoanewshelter,alwayschoos-
anemones, nearly all of large size (basal diameter >1 cm). ing a shell (P < 0.0001). In summary, 42 of the 83 (51%)
Anemones were never found on shells occupied by P. long- P. impressus captured in S. alatus shells switched into a new
icarpus and only rarely (<1%) found on shells containing shelter,showingahighlysignificantpreferenceforothershells,
C. vittatus.Anemones occurred withindividuals ofP. impres- notsponges. Results showthatindividualsofPagurusimpres-
sus and P pollicaris about 5-7% ofthe time and were com- sus, a species commonly foundin hermit crab sponges in the
monly foundon shells usedbyP. diogenes. Of39P. diogenes Dog Island area, exhibit a significant preference for shell
surveyed,26(67%)hadanemones. Ofthese,anemonenumber shelters. The likelihood with which animals switch shelters
rangedfromonetoelevenindividuals(mean=3.8). dependsonsheltertype. Intheshelterpreferencetests shelter-
TheassociationofP. impressuswithanemoneswas signifi- fidelity was also significantly higher for animals in shells
cantly less than P. diogenes and significantly greater than (31/83 or37.3%) thanforthosein sponges (19/126or 15.1%)
C. vittatus (two-sided tests for equality of two proportions: (X2= 13.63, dfl,P<0.001).
P. impressusvs C. vittatus, Z = 7.07, P<0.001 ; P. impressus Nearly all switches in the follow-up tests were long-term
vsP. diogenes,Z=7.80,P<0.001). No significantdifference and>24h.Animalsinshellsthatswitchedlong-term(24-72h),
was foundbetween P. impressus and P. pollicaris in terms of always switched into other shells, never sponges. Animals in
theirassociationwithanemones (Z= 1.50,P=0.13). spongesshowedmorevariabilityinresponseandagreaterinci-
Incidenceofstrandingonshore. Clibanariusvittatusisahardy dence ofswitching. Most animals in sponges (59%) switched
speciesthatoftenleavesthewaterandcanremainonshorefor intoashell, andthemajority (42/48 or87.5%)remainedlong-
days at a time (Rudloe, 1984). During the summer transects, term (24-72 h). Many more animals in sponges (16%, com-
manyindividualsofC.vittatus,thedominantspeciesinthebay pared to 3%) switched into both shelter choices and many of
during the summer, were found on shore. For example, in theseswitchedintoandusedallthree availablesheltersduring
the five belt transects conducted during June 1999, 1,218 thefollow-uptest.Severalanimalsswitchedfromseventonine
C. vittatusindividualswerecollected,and404(33.3%)wereon timesbetweenallthree sheltersduringa48hperiod.
shore.
Individuals of Pagurus impressus in sponge shelters are Discussion
commonly strandedon shore. However, hermitcrabs in shells
arealsooftenstrandedonshoreinJanuary,andsuchshellsare The fact that only one or two species of hermit crabs (i.e.
oftenoccupiedby aP. impressus, sometimesP.pollicaris, and PagurusimpressusintheintertidalzoneandbothP. impressus
only rarely by another hermit crab species. To determine and Paguristes hummi in the subtidal zone) commonly use
whetherindividualsofP. impressusinshellshadagreaterlike- spongesheltersintheGulfofMexicoissimilartosituationsin
lihood ofbecoming stranded than other species, the stranding otherlocationswherehermitcrabspongesoccur.InHokkaido,
of individuals of P. impressus in shells was compared with Japan,nearlyallhermitcrabspongesareoccupiedbyPagurus
P.pollicaris,theotherspeciesoftenfoundonshore.Duringthe pectinatus (Stimpson), 1858 and in southwest England and
five years, January 1997-2001, 1,124 P. pollicaris in shells Scotland by Pagurus cuanensis Thompson, 1843 (H. Mukai,
50 F. Sandford
pers.comm.; Sandford,pers. obs.).Inotherlocalitieswithher- leave sponges and enter shells occurs for Paguristes hummi
mit crab sponges, however, up to nine or more species use (Sandford, unpubl. data), the only other hermit crab found in
sponge shelters (e.g. British Columbia (Kozloff, 1987), theareathatcommonlyusessponges(Wass, 1955;Wells, 1969;
Mikawa Bay, Japan (Tanaka, 1995), and the Mediterranean Williams, 1984).
(Vosmaer, 1933). Shell/spongeswitchinghasnotbeenstudiedinsponge/crab
In nearly every location where hermit crab sponges have associations fromelsewhere, but in many locales only certain
beenreported,theyoccurindeepwaterandareonlyretrievable hermitcrabspeciesoccurasthetypicalspongeoccupants(e.g.
by scuba, trawling, dredging or other means (e.g. octopus Pagurus cuanensis and Paguristes eremita in Suberites
traps). Dog Island is unusual in that hermit crab sponges are domunculafromtheNorthAegeanSea(Voultsiadou-Koukoura
common near shore or on shore. This is partially due to the and Koukouras, 1993), Pagurus cuanensis in Suberites sp.
behaviour of P. impressus, a typically subtidal species that from the Adriatic Sea (Stachowitsch, 1980), the Irish coast
often uses sponge shelters, which migrates into the intertidal (Selbie, 1921)orsouth-westernEnglandandthewesterncoast
zone in the winter, bringing sponges closer to shore. When of Scotland (Allen, 1967; Sandford, pers. obs.), and Pagurus
animals in sponges switch into available shells, the empty pectinatusinSuberitesficusfromHokkaido, Japan(Sandford,
spongesareabandonedandeasilywashedonshore.Nearlyhalf pers. obs.). Shell preference for sponge-using hermit crabs
ofall sponges collected in the surveys were empty. Although fromtheseandotherlocationsisnotknown.
the presence ofhermit crab sponges in the intertidal zone of Wilber (1990) speculated that hermit crabs change shells
Dog Island in the winter is largely due to individuals of frequentlyandAbrams(1987),studyingshellswitchinginfive
P. impressus transporting themfromthe subtidal zone nearto hermit crab species, found that the percentage ofindividuals
shore,thenumberofemptyspongesfoundeachyearfrom1993 that switched shells ranged from 40% in Pagurus hemphilli
to 2001 did not correlate with the number of P impressus to 81% in P samuelis. In this study 47/83 P impressus
surveyed. This suggests that sponge abundance at the study (57%) switchedfromone S. alatus shellinto another, afigure
site is not due solely to the behaviour ofP. impressus but to consistentwithAbrams’ (1987)findings.
other factors, such as weather and tides, or the biology and Useofspongesas shelterscanbecostly. Spongesinterfere
reproductionofthesponges. withburyinginthesubstrate,andaremoreeasilyinfluencedby
Recentlyfield-caughtjuvenileindividualsofP impressus,a tidesandcurrentsorbeachedbywaves. OnDogIslandhermit
majorityinspongeshelters,exhibitsignificantshellpreference. crabs in sponges are seenrolling aboutinthe swashzone and
Animals in either shell or sponge shelters usually switch into areoftenwashed ashore. Thecrabs oftendiefromdesiccation
shells, notinto sponges, supportingresults ofprevious studies butthespongesrehydratesuccessfullywhenreimmersed.
on individuals maintained in the laboratory in mixed-species SinceuseofspongesheltersiscostlyandP impressusjuve-
assemblages orinconspecific groups (Sandford, 1994, 1995). niles prefer shell shelters, why are so many found occupying
A shell preference was also indicated by differences in short- sponges? Several possibilities exist. Juvenile P impressus
termversuslong-termswitching.Animalsswitchingintoshells either use sponge shelters because (i) they confer some
fromeithershellsorsponges,usuallyremainedinshellshelters survival advantage, (ii) appropriate shells are scarce, or (iii)
long-term (i.e. >5 minutes, and typically at least 24-48 h). they are out-competed by more aggressive species such as
Animals switching from one sponge into another typically Ppollicaris.
exhibitednosuchlong-termbias.Crabsinspongesshowedsig- Sponges may confer survival advantages, especially to
nificantly reduced shelter fidelity, and were more likely to smallorjuvenileindividuals ofP impressus,thetypicaloccu-
switch shelters, usually into shells. Inareas where they occur, pants. Nybakken (1996) considered the sponge/hermit crab
hermit crab sponges are used as alternative shelters by shell- association an example of true mutualism, with the sponge
dwellinghermitcrabs,althoughspongesarelesspreferredand deterring crab predators with its disagreeable taste. Taylor
arelikelysuboptimalshelters. (1994) noted that shell-encrusting bryozoans may reduce pre-
These results contradict the suggestion (Benedict, 1900; dationbecause ofshell thickening, chamberenlargement, and
Vosmaer, 1933; Rabaud, 1937; Hart, 1971) that one possible camouflage.Thesesameargumentscouldapplyforshell-over-
benefitofuseofaspongeshelterbyahermitcrabistheadvan- growing sponges, and there is some evidence to suggest that
tage ofalongertermassociation withaliving, growinghome hermitcrabs in sponges are less likely tobe attackedoreaten
andareducedneedtochangeshelterscomparedtohermitcrabs by blue crabs, Callinectes sapidus, a typical predator, either
in shells. Encrusting sponges andbryozoans are the only gas- because ofa chemical, texture, or camouflage factor (Farley,
tropod shell epibionts associated with hermit crabs which pers.comm.; Sandford, pers. obs.).
wouldallowthecrabtoretainashellshelterforalongerperi- In laboratory conditions, use of sponges by individuals of
odbyconstantlygrowingandeffectivelyenlargingthevolume P. impressus increases as shells become less available
ofthe shelter (Stachowitsch, 1980; Taylor, 1994).Although it (Sandford, 1995)sointhefieldcrabsmayusespongesinthose
hasbeensuggestedthatencrusting symbiontsgrowingongas- habitats where shells are relatively scarce. Shell scarcity is a
tropod shells may reduce the frequency of switching, Taylor limitingfactorforhermitcrabs (Kellogg, 1976, 1977; Hazlett,
(1994)notedthatthishasneverbeendocumentedbylong-term 1981; Pace, 1993) and may be a major factor explaining the
observations of individual hermit crabs. This study confirms largenumberofP impressususing sponge sheltersintheDog
that individuals ofP. impressus using sponge shelters do not Islandarea.
typicallyexhibitlong-termattachments.Asimilartendencyto Shelluseandaggressioninhermitcrabsarelinked(Hazlett,
Populationdynamicsandepibiontassociationsofhermitcrabs 51
1981) and it has been suggested (Rudloe, 1999) and there is habitatand smallersizerelativetotheotherspecies examined
some evidence (Kellogg, 1977; Hazlett, 1980; Sandford, pers. for anemones, likely explains the absence of anemones on
obs.)thatindividualsofP. impressusarelessaggressiveorless shellsoccupiedbythisspecies.IndividualsofPpollicarisalso
competitive than other hermit crab species. Sympatric hermit commonly bury in the substrate (McLean, 1974), commonly
crab species can coexist in spite of shell competition by associatewithC. tricolor(Rudloe, 1999),andwereoftenfound
exhibiting different shell preferences, habitat preferences, or with anemones in this study, so their greater association with
otherbiologicaldifferences(Hazlett, 1981). Kellogg (1977)in theanemonerelativetoP. longicarpusismostlikelyexplained
Beaufort Harbor, North Carolina, studied many of the same bytheirgreaterexposuretohabitatsinthesubtidalzonewhere
hermitcrab species asinthis study. Hefoundthathermitcrab anemonesmorecommonlyoccur.
populationstherewereshell-limitedandthatbothP. impressus ApreferentialassociationofH. echinatawithcertainhermit
and P. hummi individuals were infrequently found in inshore crabspecieswasreportedbyYundandParker(1989).Thecom-
areas, and were likely to be outcompeted by P. pollicaris pleteabsenceofH. echinataonshellsusedby C. vittatusmay
and P. longicarpus. In the Dog Island area individuals beexplainedbythisspecies’preferencefortheupperintertidal
of both P impressus and P hummi may use the sponges zoneand/oritstendencytospendtimeonshore.Thetotallack
as less favoured alternative shelters as a consequence ofsuch ofassociation ofthe hydroid withP. impressus is mostlikely
competition. duetotherejectionofhydroid-coveredshellsbyindividualsof
The possible benefits of sponge shelters remain undeter- this species.
mined,butitisnowclearthat spongesare suboptimal shelters
that may affect locomotion and increase mortality risks Acknowledgements
from dehydration due to stranding. Although many juvenile
eaxntdensdoemdepeardiuoldts,inmdoisvtidsuwailtschofinPt.oishmeplrlesswsuhsenusaevaislpaobnlge.es for ImeamingmrayteffiuelldtowotrhekmoannDyogCoIeslCanodlleeagcehsJtaundueanrtysfwrhooma1s9s9i3stetod
Forthe othertwo epibionts studied, Calliactis tricolorand 2001,toGavinCrossandPeggyKnottfortheirhelp,toSharon
Hydractinia echinata, the results clearly show a significant Sandfordforheradvice and support, andtoDanRittschoffor
association ofP pollicaris with the hydroid H. echinata and his helpful comments on an early draft ofthe ms. I am espe-
a significant lack of association of both P. impressus and cially indebted to Guest Editors, Rafael Lemaitre and Chris
C. vittatus with the same hydroid. The cloak anemone Tudge, and three anonymous reviewers for their critically
Calliactis was associated with three species (P pollicaris, usefulcommentsonthefinaldraft.
P impressus, and especially P diogenes) and not associated
withtwoothers {P. longicarpus and C. vittatus). Thecommon
associationofCalliactistricolorwithP diogenesfoundinthis References
study has been noted by Rudloe (1984) who has found
P. diogenesindividualswithasmanyas20anemones. Abrbaemtsw,eePn.Ai.nt1e9rt8i7d.alRehesromuirtcecpraarbtsitoinontihnegoauntdercocomapsettiotfioWnasfhoirngstheolnl.s
The association of anemones with P. impressus, Oecologia72:248-258.
Ppollicaris,andP diogenesistobeexpected,astheyfrequent Allen, J.A. 1967. Thefauna ofthe Clyde Sea area - Crustacea:
subtidal hard bottom habitats more optimal for anemones. Euphausiacea and Decapoda. Scottish Marine Biological
Conversely, the two intertidal species, P. longicarpus and Association:Millport,Scotland.
C. vittatus, that live in sandy bottom habitats which are less Benedict, J.E. 1900. The Anomuran collections made by the Fish
optimal for anemones, lacked anemones. The deficiency of Hawk expedition to Puerto Rico. Bulletin ofthe U.S. Fisheries
anemonesonshellsoccupiedbyC. vittatusmayalsoberelated Commission20: 129-148.
to the fact that C. vittatus is the hardiest of all hermit crab Brooks,W.R., Ceperley,L., andRittschof, D. 1995. Disturbance and
species in the GulfofMexico, and oftenleaves the waterfor reattachment behavior of sea anemones Calliactis tricolor (Le
Sueur): Temporal, textural, and chemical mediation. Journal of
daysatatime(Rudloe, 1984, 1999).Undersuchcircumstances, ChemicalEcology21(1): 1-12.
anyanemonescarriedontheshellswoulddesiccate.Anyasso- Hart, J.F.L. 1971. New distribution records ofreptant decapod Crus-
ciationofC. vittatuswithC. tricolorisnon-symbioticinnature tacea,includingdescriptionsofthreenewspeciesofPagurus,from
(Brooks et al., 1995). Individuals of Calliactis tricolor are the waters adjacent to British Columbia. Journal Fisheries
foundrarely attachedto shellsoccupiedby C. vittatusbutitis ResearchBoardofCanada28(10): 1527-1544.
notcertain whetheranemones do not discriminate (Brooks et. Hazlett, B. 1980. Communication andmutual resource exchange in
al., 1995; Sandford, pers. obs.) oractuallyavoid C. vittatusas NorthFloridahermitcrabs. BehavioralEcologyandSociobiology
reportedbyRudloe(1984). 6: 177-184.
ThelackofassociationofP longicarpuswiththeanemone Hazlett, B. 1981. The behavioral ecology of hermit crabs. Annual
may also be due to this hermit crab’s burying behaviour. ReviewofEcologyand Systematics12: 1-22.
Kellogg,C.W. 1976. Gasttopodshells: apotentiallylimitingresource
Althoughfeatures suchas shell sizeorshelldamageaffectthe for hermit crabs. Journal ofExperimental Marine Biology and
switching behaviour and survival of species like P pollicaris Ecology22: 101-111.
(McClintock, 1985), Kuhlmann (1992) concluded that for P Kellogg, C.W. 1977. Coexistenceinahemhtcrab speciesensemble.
longicarpus burying behaviour is probably more important BiologicalBulletin 153: 133-144.
than shell features as an anti-predator factor. The burying Kozloff, E.N. 1987. Marine invertebrates ofthe Pacific Northwest.
behaviour ofP longicarpus, in addition to its intertidal zone UniversityofWashington Press: SeattleandLondon.
52 F. Sandford
Kuhlmann,M.L. 1992.Behavioralavoidanceofpredationinaninter- Stachowitsch,M. 1980.Theepibioticandendolithicspeciesassociat-
tidal hermit crab. Journal ofExperimental Marine Biology and ed with the gastropod shells inhabited by the hermit crabs
Ecology157: 143-158. Paguristes oculatus and Pagurus cuanensis. Publicazioni della
McCormack,G.P.,andKelly,M. 2002.Newindicationsofthephylo- StazioneZoologicadiNapoli:MarineEcology 1:73-101.
genetic affinity of Spongosorites suberitoides Diaz et al., 1993 Strasser,K.M.,andPrice,W.W. 1999.Anannotatedchecklistandkey
(Porifera, Demospongiae) as revealed by 28S ribosomal DNA. to hermitcrabs ofTampa Bay, Elorida, and surrounding waters.
JournalofNaturalHistory36: 1009-1021. GulfResearchReports, 11:33-50
McClintock,T.S. 1985.Effectsofshellconditionandsizeuponshell Tanaka, T. 1995. Studies ofa sponge. Suberitesficus (Hadromerina,
choicebehaviorofahermitcrab.JournalofExperimentalMarine Demospongiae), living in symbiosis with hermit crabs and a sea
BiologyandEcology88:271-285. snail, Synthopsis spongicola, inMikawaBay. EacultyJournalof
McLean,R.B. 1974.Directshellacquisitionbyhermitcrabsfromgas- AichiGakuinJuniorCollege3: 112-128.
tropods.Experimentia30: 206-208. Taylor,P.D. 1994.Evolutionarypalaeoecologyofsymbiosesbetween
Nybakken, JW.Wm. 1996. Diversity ofthe invertebrates - a laboratory bryozoansandhermitcrabs.HistoricalBiology9: 157-205.
manual. C.BrownPublishers:Dubuque,Iowa. Yund,P.O.,andParker,H.M.1989.Populationstructureofthecolonial
Rabaud,E. 1937.Notes sonunaires surlesrapportsdespaguresavec hydroidHydractinia sp. nov. C inthe GulfofMaine. Journalof
les actinies et les eponges. Bulletin de la Societe Zoologique de ExperimentalMarineBiologyandEcology125:63-82.
Erance62:400-406. Vosmaer, G.C.J. 1933. The spongesoftheBayofNaples-Porifera
Rudloe,J. 1984.Theeroticocean. NewYork:E.P.Dutton,Inc. Incalcaria, Vol.I.MartinusNijhoff:TheHague.
Rudloe, J. 1999. GulfSpecimen Marine Laboratories Inc. catalog. Voultsiadou-Koukoura, E., and Koukouras, A. 1993. Remarks on
GulfSpecimenMarineLaboratories:Panacea,Elorida. sponge-decapod associations in the North Aegean Sea. Bios
Sandford, P. 1994. The Elorida hermit-crab sponge, a little known (Macedonia, Greece) 1(1):252-256.
“mobile”spongefromtheNEcorneroftheGulfofMexico,andits Wass, M.L. 1955. The decapod crustaceans ofAlligatorHarbor and
hermitcrabassociations.Pp. 273-278in: vanSoest,R.W.M.,Van adjacentinshoreareasofNorthwesternElorida. QuarterlyJournal
Kempen,T.M.G.,andJ.C.Braekman(eds).ProceedingsoftheTVth oftheEloridaAcademyofSciences18(3): 129-176.
International Porifera Congress, Amsterdam, the Netherlands, Wells, H.W. 1969. Hydroid and sponge conunensals of Cantharus
April 19-23, 1993.Balkema:Rotterdam. cancellariuswitha“falseshell”.Nautilus82(3):93-102.
Sandford, E. 1995. Shell/sponge switching by hermitcrabs, Pagurus Wilber,TP 1990.Associationsbetweencrabmorphologyandgastro-
impressus.InvertebrateBiology4(1):73-78. podshellspeciesinthehermitcrabPaguruslongicarpus. Journal
Sandford, E, andBrown, C. 1997. Gastropod shell substrates ofthe ofCrustaceanBiology 10(1):134-138.
Eloridahermit-crab sponge, Spongosorites suberitoides, fromthe Wilber,P.1993.Pseudoreplicationinhermitcrabshellselectionexper-
GulfofMexico.BulletinofMarineScience61(2):215-223. iments:doesitoccur?BulletinofMarineScience52(2): 838-841.
Sandford,E,andKelly-Borges,M. 1997.Redescriptionofthehermit- Williams,A.B. 1984.Shrimps,lobsters,andcrabsoftheAtlanticcoast
crab sponge Spongosorites suberitoides Diaz, Pomponi and Van of the eastern United States, Maine to Elorida. Smithsonian
Soest.JournalofNaturalHistory31:315-328. InstitutionPress:Washington,D.C.
Selbie,C.M. 1921. TheDecapodaReptantiaofthecoastsofIreland,
PartII:Paguridea.Eisheries,Ireland,ScientificInvestigations,No.
I.HisMajesty’sStationaryOffice:Dublin.